conversion of extremely saturated photos to cmyk

[girlsfather]

Here is a question to all the preprint colour-experts:
What is the best way to convert saturated pictures that show aereas that do not fit in the cmyk colourspace from rgb to cmyk ?

I have a very saturated photograph, 16 bit in eci-rgb_v2.icc.
When converting to cmyk, isocoated_v2.icc, in CS3 the background looks awful (colour clipping in the soft colour-gradients
I tried all methods and relative colorimetric is the most pleasant.

The best I found out so far is: Reduce Saturation in CS3 by 40%, that is the amount necessary to avoid clipping.
Then convert to profile.
Then in cmyk increase saturation again.
The attached Photo shows part of the background.

The gradient background keeps smooth with the -40/+40 saturation method.
The colour-change is not really nice.
Is there a better, more professional way?

Thank you for every help!!!

Michael

[TommyO]

Michael,
Will take a look at it. Also, could you post images (or zip) we can actually play with, in another format... the jpeg's get converted to sRGB and loose all relative differences in color. So, it does not do much good to look at those. It would also be better if they were separated, not in one image.

Typically, it is best if you control the conversion/desaturation and not the conversion engine. At least that way you some control over the result. What you're doing is fine, as long as you like the result. Base final result on a hard proof though.

[pixelzombie]

i have also found that reducing the saturation before your conversion will make a huge improvement in your cmyk file...

[TommyO]

After leaving and thinking, it dawned on me that you really did not describe what your workflow or end result was targeting, i.e. are you trying to profile/match your contract proofer or your press? ...what paper stock? ...is there a color server that can do it for you?

Since presses vary in their output and are expensive to profile, many people profile the contract proofer. It is also generally more successful matching the contract proofer. The colorants used by the proofer are similar to press inks in their dot gain characteristics, etc. Then you end up with separations that work for both, making your workflow easier.

The benefit is being able to set limits on ink and black generation and create separations that ready for the press. Of course, if there is a color server that can do the profile conversions from press to proofer for you, that is ideal. It can make adjustments for the changes in dot gain on differing paper stocks and apply the transfer characteristics for the platesetter or imagesetter. The resulting profile then becomes your aim point for your RGB-to-CMYK conversions and can serve as your proof profile as well.

[girlsfather]

[quote=TommyO;207217]After leaving and thinking, it dawned on me that you really did not describe what your workflow or end result was targeting, i.e. are you trying to profile/match your contract proofer or your press? ...what paper stock? ...is there a color server that can do it for you?

Hi Tommy,

thank you very much for your answers and questions.
I am a professional photographer, doing my own retouching.
I work in Photoshop CS3 with calibrated TFT and custom printer profiles (RGB) for my Epson 4800 to print my portfolio.

This Image is going to be 4c offset-printed for selfpromotion in an exclusive artbook-publication.
My aim is to see the possible result in advance.
So I checked the Numbers and Photoshop tells me that great parts of especially the background is not within the printable colourspace of isocoated_v2.

I normally deliver both - RGB-Files and the converted CMYK.
In this case I have the impression that the Contractproof and later the offsetprint will look awful.
Maybe I am totally wrong?

I am looking to talk to the printing house and their experts.
I want to learn more about conversion from RGB to CMYK, so I posted this question here.

So, do I need to own other applications than photoshop to do the conversion?

Here now a jpg with maximum quality and the eci-rgb_v2 embedded.
Still quite small pic - I do not have a possibility to put a bigger one online.
(This limitation to 100KB seems set a bit too low ;-)))

[pixelzombie]

is your epson calibrated for the presshouse that will be printing this book?

[girlsfather]

Originally Posted by pixelzombie

is your epson calibrated for the presshouse that will be printing this book?

Hi Pixelzombie,

no the printer is profiled to give the best result printing in RGB only.
Only for my portfolio-book I show to the clients - not for proofing.
I do not have a RIP software yet.

I have a lab that does GMG Contractproofs for me.
The lab wants CMYK-Files...

I want to see in advance what would be the best possible result.
I am aware that this will not be 100% what i will get later in offsetprinting.
Just want to know how it is possible to have that file printed closest to the RGB monitor softproof.......

[pixelzombie]

you could have the presshouse run the book with higher densities, you could also run a touchplate on that image which would mean a 5th color on press(more $) and the other alternative is hexachrome printing...

[Damo77]

Originally Posted by girlsfather

When converting to cmyk, isocoated_v2.icc

Have you been told this is definitely the most accurate profile for the press? If so, do you believe it?

This thread sounds like another one for my "why large-gamut RGB colour spaces are pointless" file ...

[girlsfather]

Originally Posted by Damo77

Have you been told this is definitely the most accurate profile for the press? If so, do you believe it?

This thread sounds like another one for my "why large-gamut RGB colour spaces are pointless" file ...

That is the profile all the stuff is contractproofed here in germany.
The profile was created by the European Color Initiative (ECI)
as ISO 12647-2 was redefined.
The profiles are based on Fogra39L.
These profiles are used since last year and replace isocoated.icc.

[Damo77]

Hmmm, that may or may not answer my question.

The way I see it is, you have two choices - (a) the brightest possible colour, which will almost certainly not be accurate, or (b) the most accurate possible colour, which will almost certainly not be the brightest the press can produce. In most cases, we pursue choice (b) - that's the nature of colour management. However, I have certainly chosen to pursue choice (a) for various jobs in the past.

If you pursue choice (b), then you need to know that the ICC profile you're using is accurate - hence my original question. Use your soft-proofing and gamut warning, and massage your o-o-g colours into the destination gamut. By the way, I'm not sure about your global 40% desaturation, I'd prefer to see you desaturating particular colours as needed. In Photoshop's Hue/Sat dialog, use the Sat slider to desaturate each colour enough to get rid of those "!" marks in your Info palette, and where necessary, nudge the Hue slider to keep the colour as true as possible to the original. Just like TommyO said, it's better to do this yourself, and keep exquisite control over it, than let an automatic process control it for you.

If you pursue choice (a) - that is, saturation at all costs, regardless of colour accuracy - then I would still do the careful desaturation as described above, then convert, then do whatever it takes to get your colours to their max (eg reds 100M100Y, blues 100C100M etc). This might mean Curves, or Levels, or Selective Color, or Hue/Sat (although I've found Hue/Sat to be awkward in these cases). The point is, once you start trying for those extreme CMYK colours, you'll be departing from the colours in your original file.

[girlsfather]

TommyO, pixelzombie, Damo77,

thanks to you all!
I will play around with the picture and will see what turns out to be the best.
At the moment there are some other, more urgent things to do....

P.S. I tried before to desat/change only specific colours in the picture. But especially in the colour gradients of the background this really is not easy!